Am 09.03.2018 um 15:05 schrieb Michel Dänzer: > On 2018-03-09 02:46 PM, Christian König wrote: >> We unmapped imported DMA-bufs when the GEM handle was dropped, not when the >> hardware was done with the buffere. > Good catch! I guess that explains some oddities I sometimes saw with > PRIME. Both patches are Yeah and it took me a month to hunt that down. Thanks to IOMMU or otherwise I wouldn't have even noticed that something is wrong. > Reviewed-by: Michel Dänzer <michel.daenzer at amd.com> Thanks, Christian.